Deloitte China Monthly Report Issue 99
Economy
Boosting consumption to contain involution
Will growth in the second half be a letdown after a strong showing in the first half of 2025 with growth of 5.3% yoy? Are tariffs losing their relevance? Have financial markets, which have shrugged off recent crude tariff threats from the Trump Administration, underestimated the potential economic risks resulting from a fragmented global trading system? Should China step up its export diversification away from the US, given that China faces higher tariffs than other US trading nations? What are the chief causes of involution? What could be done to address the issues without compromising economic growth?
For the second half of 2025, it will be difficult to replicate the strong GDP growth seen in 1H, mainly due to renewed weakness in the housing sector, despite some signs of stabilization in May, and greater efforts at anti-involution (meaning to reduce production). We still believe that the 2025 growth target of “around 5%” remains on track. Based on various surveys of the housing sector, existing home prices' decline has become more widespread, affecting even first tier cities in July. As housing prices continue to trend lower, more potential buyers are staying on the sidelines. As a result, the knock-on effect on housing-related merchandise is weighing on consumption. The housing sector’s challenges are contributing to existing deflationary pressures, with July CPI YoY growth at 0%. Price wars in many sectors have led to an official campaign on anti-involution which covers traditional industrial sectors and extends to services such as delivery. However, the reasons for involution are not merely due to cyclical downturns. Overcapacities have existed in many sectors, but the housing boom in previous years has masked such imbalances. The escalating tariff conflicts and shifting geopolitical landscapes have even heightened the vulnerability of industries with overcapacities to various trade barriers and restrictions, such as President Trump's adamant stance on steel tariffs on the grounds of national security.
Chart: Bank loans have seen contraction due to consumer cautions
Source: Wind
Meanwhile, exports have continued to strengthen, posting a 7.2% yoy increase in July, despite the ongoing shift away from the US market. The contribution of the US market as an export destination has declined since 2018 (from 19.2% in 2018 to 12.0% in the YTD period ending 2025 May). We attribute the external sector’s stellar performance to export frontloading as well as China’s competitiveness. However, such competitiveness has also stoked protectionism, which is partially driven by fears over China’s alleged overcapacities. At some point, the export front-loading effect will subside. Our baseline scenario is that the economy may experience a mild letdown in the second half of the year, and a revival of the property sector is unlikely. Regarding reflation-related policy measures, unless trade disruptions bring about severe external shocks to growth, we do not foresee a large fiscal stimulus.
Is the tariff war over?
Following a few months of impasse, the Trump Administration has wrapped up most trade deals before the August 1 deadline. The conventional wisdom is that the US has effectively cornered major trading nations into a “prisoner’s dilemma”, leaving trading partners with little choice but to offer concessions. Broadly speaking, the developed world has signed up for 15% tariffs on their exports to the US along with investment commitments (e.g., Japan and South Korea). Australia and Singapore have agreed to a 10% tariff, despite running a trade deficit with the US. Most ASEAN countries, on the other hand, are subject to a tariff rate of around 20% or higher, depending on the extent of perceived transshipments from China. It is almost certain that tariffs for China, even with a trade deal with the US, will be higher than those imposed on the rest of Asia.
Market reaction and China’s policy responses
Since the initial declines of capital markets following the tariff announcement on April 2, the markets have bounced back strongly. The US stock market has rallied significantly since mid-April to nearly all-time highs recently. This rebound assumes that tariffs would not be excessively high, and that some tariffs might encounter legal hurdles in the US, potentially leading to their reversals. Is such expectation excessively optimistic? Recent US economic data, such as the July 2025 non-farm payrolls growth (totaling 73,000) and Q2 GDP growth of 3% (seasonally adjusted annual rate), suggest potential economic weakness caused by delayed investment. In other words, the impact of high tariffs has not yet shown up in the US economy.
Meanwhile, investors remain convinced that AI-led investments will continue to boost corporate earnings. As such, macroeconomic factors have temporarily taken a backseat. For the Asia Pacific region, the tariffs represent a managed setback, and the overall competitiveness and growth expectations remain strong. The timing of a trade deal between the US and China depends on numerous variables, including the potential outcomes of the Trump-Putin meeting on August 15. So far, neither a cease fire nor a peace treaty has been on the cards in the wake of Trump-Putin summit. Such stalemate on the battlefield is mildly negative for crude oil prices. For China and rest of Asia, central banks would have had more reasons to ease.
On its trade negotiations with China, the US appears to lack a clear objective. Unlike the trade war 1.0, when reducing trade deficits was the overarching objective, China is now viewed by the US as a major rival in manufacturing, technology, and military capacity. In this context, it is natural for China to perceive US call of “rebalancing the economy” as a ploy to hinder its development. Until both nations come up with clearly defined rules on export controls and investment provisions, the most likely scenario is an extended pause in tariff escalations.
The most closely watched financial variable is long-term interest rates, as the Fed is leaning to a more dovish stance. However, if the Fed’s easing campaign is seen as politically motivated, the Treasury market may react negatively by pushing long-term interest rates higher. The Fed’s easing is likely to trigger another round of rate cuts in Asia, including in China. But with trade tensions between China and the US having eased somewhat, the sense of urgency for the People’s Bank of China to act may be reduced, unlike the rate cut observed in April.
Chart: Confidence returns to A-share market on hopes of reflation
Source: Wind
The 10% universal tariffs are likely to remain in place for a long time. Again, it could have been worse. Some countries like Canada and India have taken a page from China by standing firm or eventually offering limited concessions. As such, the global trading system will become more fragmented. In the medium term, most countries will have to seek alternatives. For China, shifting to domestic demand-led growth will become more imperative before export growth momentum peters out. The policy bias toward consumption over infrastructure is also about avoiding further involution, as previous fiscal stimulus programs have all increased capacities in industrial sectors. On consumption-boosting measures, we expect the cash-for-clunkers program to continue in order to incentivize appliance purchases. Would such a scheme extend to housing? There have been signs that restrictions on housing demand are being relaxed, albeit at a very slow pace (e.g., restrictions were lifted in outskirts in both Beijing and Shanghai). This is a welcome move. We continue to advocate for “coupons” for new home renovations, as improved demand for housing will result in a much bigger bang for the buck.
We will expect more policy directives aimed at preventing excessive price reductions. In sectors where overcapacities are severe (e.g., steel and cement), production reductions will likely be called for. Ultimately, consolidation of these sectors will have to take place. To make the transition less painful, export growth must be sustained, especially in overseas markets where geopolitical risks do not present a binding constraint; this applies to steel exports in the ASEAN region and the Middle East. Of course, a more effective solution is to prevent abuses of industrial policies and duplicated efforts in creating champions by local governments. The fourth Plenum in October is expected to shed light on the central government’s calibrated priorities in strategic sectors.
TMT
AI: Innovation and governance in tandem
The 2025 World Artificial Intelligence Conference (WAIC), held in Shanghai marked a significant turning point in the trajectory of global AI development. More than a showcase of cutting-edge technology, the event underscored China's evolving ambitions in AI leadership—not only as a technology hub but also as a governance architect and global collaborator. We have identified five critical trends observed at WAIC 2025, assesses their underlying strategic intent, and provides recommendations for stakeholders navigating this evolving AI ecosystem.
1. From language models to embodied intelligence
The seamless integration of large language models (LLMs) with physical agents is ushering in a new era of embodied intelligence. Robots are no longer presented as novelties or prototypes; rather, they are positioned as viable, near-term solutions for logistics, manufacturing, eldercare, and even retail. This shift reflects more than a technological evolution—it represents a rethinking of how intelligence operates in physical space. By embedding cognition into mobile, sensor-rich platforms, Chinese developers are seeking to reengineer service sectors and automate blue-collar roles with contextual reasoning and interactive fluency. The most advanced humanoid robots demonstrated multimodal capabilities—responding to voice, adapting to movement, and performing fine-motor tasks once considered out of reach for machines. For companies, this signifies a maturing investment landscape, where LLM-augmented robotics are moving out of research labs and into commercial pilots. Firms in manufacturing, healthcare, and logistics should view embodied AI not as a futuristic concept, but as a next-generation interface capable of absorbing routine human tasks in real-world settings.
2. Industrial AI integration beyond pilots
The shift is clear: AI has become not just an enhancer, but an orchestrator of industrial activity. What distinguishes the current wave of industrial AI is the depth of its integration. Rather than isolated deployments, companies are now threading AI through ERP systems, IoT networks, and digital twins. The result is a more responsive, resilient, and intelligent manufacturing environment—capable of anticipating disruptions, autonomously reallocating resources, and continuously optimizing workflows. The strategic opportunity lies in rethinking value creation. Instead of treating AI as an overlay, forward-thinking organizations should embed it into the design of core workflows. This requires more than technical upgrades; it demands a transformation in how human labor is organized, how systems interact, and how data is interpreted. For private-sector stakeholders, the imperative is to move from experimentation to orchestration—ensuring that AI is not only integrated, but also mission-aligned.
3. Rise of sovereign AI and computing infrastructure
Geopolitical dynamics are reshaping AI infrastructure strategies. Enterprises and governments alike are pursuing sovereign AI frameworks to ensure control over foundational models, compute resources, and data governance. This includes localized large language models trained on domain-specific datasets, along with increased investment in domestic GPU and AI chip manufacturing. The emerging consensus is that without ownership of the full AI stack—hardware, algorithms, and data pipelines—long-term competitiveness cannot be assured. Organizations are now evaluating supply chain dependencies and investing in multi-tier resilience strategies that balance cloud, edge, and on-premises computing.
4. Open-source AI models proliferate
By positioning itself as a supplier of high-quality, license-free AI models, China is accelerating adoption in regions with limited access to proprietary Western systems. This approach not only fosters goodwill but also creates strategic dependencies on Chinese-developed toolchains, model architectures, and support ecosystems. For global companies, the implications are significant. Open-source models may offer a faster and more cost-effective path to LLM integration, particularly for domain-specific applications. However, they also require robust internal governance to ensure quality assurance, bias mitigation, and compliance with varying national regulations. Firms must weigh openness against oversight and evaluate the sustainability of open ecosystems amid rapidly evolving regulatory environments.
5. Global AI Governance is Shifting
WAICO is envisioned as a multilateral body tasked with establishing shared standards, safety protocols, and cross-border governance norms for AI systems. The WAICO proposal repositions China not merely as a technology exporter, but as a convening power for global AI cooperation. The model invites participation from the Global South and developing economies—many of which are underrepresented in forums like the OECD or G7. By anchoring governance in consensus and openness, China is offering a model that resonates in a multipolar world. This pivot carries long-term strategic consequences. Regulatory divergence is likely to deepen, creating parallel ecosystems of governance, compliance, and innovation. Multinational corporations must prepare for a world in which AI systems are not universally portable. Data handling practices, algorithmic transparency requirements, and ethical benchmarks may soon vary widely depending on geopolitical alignment. Strategic planning must therefore include region-specific compliance architectures and active participation in shaping emerging standards.
Strategic implications for enterprises
Industry leaders are no longer competing solely on technology access; they are competing on speed of integration, infrastructure resilience, and the ability to align AI strategies with regulatory and societal expectations. For Chinese enterprises, the emphasis on sovereign AI adds an additional strategic layer—requiring capital investment, talent acquisition, and ecosystem partnerships to secure the full value chain.
Strategic Priorities:
- Prioritize operational integration over experimentation — Focus on end-to-end deployment pipelines that move rapidly from prototype to production.
- Invest in infrastructure sovereignty — Develop in-house capabilities or domestic partnerships for critical AI hardware, models, and data platforms.
- Embed sustainability as a competitive advantage — Use AI to optimize energy consumption and resource efficiency in core operations.
- Build domain-specialized capabilities — Fine-tune models for the specific contexts in which the organization competes.
- Rethink workforce strategies — Integrate AI literacy across all organizational levels to unlock human-AI collaboration.
- Institutionalize governance frameworks — Ensure compliance, ethics, and risk management are built into AI system design from day one.
In conclusion, artificial intelligence in 2025 has moved beyond the hype curve and into a decisive phase of industrial and societal integration. The conversations underscored that AI leadership will not be determined by the size of one’s model or dataset alone, but by the ability to orchestrate a secure, efficient, and ethically governed ecosystem that delivers measurable outcomes. The winners of this next AI wave will be those who combine technical excellence with operational discipline, regulatory foresight, and strategic adaptability.
Logistics
Takeout war reshapes instant logistics
China’s instant logistics landscape is undergoing profound transformation amid the fierce “takeout war”.
Takeout war: disrupting the Industry
In February 2025, JD.com entered the food delivery market with a policy of “annual commission waiver + CNY 10 billion in subsidies + full social insurance for riders,” disrupting the previously stable instant logistics sector. This move prompted strong responses from Meituan and Ele.me, leading to escalating subsidy and traffic competition among platforms. From April to June, the three major platforms collectively invested tens of billions of yuan in subsidies, with single-day order volume peaking at 250 million—marking a 150% increase compared to the beginning of the year.
While the market expanded rapidly, it also exposed multiple issues on both the supply and demand sides, drawing regulatory attention. On the supply side, small and medium-sized restaurant brands lacked effective contingency plans, losing nearly one-third of their revenue during order surges. On the platform side, the three major players collectively incurred losses exceeding CNY 10 billion. Notably, JD.com reported a net profit attributable to shareholders of CNY 6.2 billion in the first half of 2025—a 51% year-on-year decline.
On July 18, the State Administration for Market Regulation (SAMR) summoned the three platforms, urging them to cease irrational competition. By August 1, all three had simultaneously issued statements halting extreme subsidies and pledging to standardize promotional activities. Thus, the six-month takeout war temporarily concluded under regulatory intervention, ushering in a phase of rational competition. This competition not only reshaped market dynamics between supply and demand but also established a regulatory framework—laying the institutional foundation for the industry’s long-term structure. In February 2025, JD.com entered the food delivery market with a policy of “annual commission waiver + CNY 10 billion in subsidies + full social insurance for riders,” disrupting the previously stable instant logistics sector. This move prompted strong responses from Meituan and Ele.me, leading to escalating subsidy and traffic competition among platforms. From April to June, the three major platforms collectively invested tens of billions of yuan in subsidies, with single-day order volume peaking at 250 million—marking a 150% increase compared to the beginning of the year.
Escalating competition: catalyzing new transformations
While the six-month takeout war has temporarily concluded, its impact on China’s instant logistics industry has been far-reaching—catalyzing a series of transformative changes in both market scope and operational models.
1. Expanding the boundaries of instant retail and broadening the market scale of market scale
Through aggressive subsidies, platforms have extended user demand from catering to daily necessities, marking a shift from “delivering meals” to “delivering almost everything.” In terms of category share, food delivery orders decreased from 68% in 2024 to 55%, while non-food categories—such as fresh produce, supermarkets, pharmaceuticals, and electronics—increased to 45% (according to Meituan’s 618 Campaign Report, 2024).
The order structure has also diversified: JD.com’s Hourly Purchase partnered with 3C digital stores to enable one-hour iPhone deliveries; Meituan Instant Shopping collaborated with Yonghui Supermarkets to provide daily groceries; and Ele.me integrated with Tmall Supermarket for hourly delivery services. This evolution has expanded the instant retail ecosystem, not only reshaping the competitive landscape but also normalizing the concept of “delivering almost everything via food delivery platforms” for consumers.
2. Optimizing resource allocation and advancing employment compliance
In response to capacity and timeliness challenges brought on by the surge in orders, the three major instant logistics platforms pursued improvements across three key areas: rider dispatch, supermarket collaboration, and technological transformation.
Rider dispatch: Platforms improved job attractiveness by enhancing employee benefits and safeguarding labor rights. JD.com led the way by offering full social insurance and housing provident funds for full-time riders. Meituan allocated CNY 1.5 billion in insurance premiums for 7 million riders, reducing barriers to insurance enrollment. Ele.me leveraged technology to increase algorithm transparency, thereby improving income fairness and dispatch efficiency.
Supermarket collaboration: Platforms accelerated the integration of domestic warehouse and distribution networks with supermarkets. JD Logistics, for example, deepened its expertise in warehouse management and supply chain optimization to streamline domestic supermarket operations. Simultaneously, it expanded international warehouse partnerships to enhance domestic instant retail efficiency through global supply chain synergies.
Technological transformation: AI applications played a key role in improving market governance. Ele.me deployed an AI language assistant to simplify merchant onboarding and implemented its “Holographic Shield” system to detect infringing products, non-compliant items, and fraudulent stores. These technologies strengthened compliance management, improved platform reliability, and promoted healthier competitive norms across the industry.
Future outlook: rational return of the value chain
The 2025 takeout war marked a turning point for the instant logistics market. Over the past five years, China’s instant delivery industry has experienced rapid growth, with a compound annual growth rate (CAGR) of approximately 19% from 2020 to 2024, reaching a market size of CNY 417.7 billion in 2024 (iResearch data).
Now, with regulatory intervention and value chain restructuring, the industry is shifting from short-term, price-driven competition to efficiency- and compliance-driven high-quality development. Deloitte forecasts a steady CAGR of around 14% over the next three years, signaling a maturing sector anchored in long-term sustainability.
Figure: Scale of China’s instant logistics market (in Billions of CNY)
Source: iResearch, Deloitte Research
Following regulatory intervention to halt extreme subsidies, the business logic of the instant retail industry has undergone a dual transformation: a shift from “exchanging subsidies for volume” to “efficiency- and structure-driven growth,” and a transition from “short-term traffic spikes” to “long-term fulfillment quality and supply chain optimization.” This evolution has directly prompted major players in the instant logistics market to recalibrate their strategies, with both platform operators and regulators adapting accordingly.
Meituan: Focuses on consolidating its dominance in food delivery while expanding into non-food categories, such as daily necessities and pharmaceuticals, to maintain leadership in fulfillment density and timeliness. It is also accelerating global expansion by entering markets like Hong Kong and the Middle East through its Keeta app.
Alibaba: Continues to leverage the Taobao Flash Sale portal to increase user engagement by boosting browsing frequency and click-through rates on the Taobao app. It aims to create synergies between far-field e-commerce (traditional cross-regional e-commerce) and near-field e-commerce (local instant retail). At the same time, it is driving operational efficiency through AI innovations to address key challenges and expand ecosystem partnerships.
JD.com: Is capitalizing on internal consumer research (source: JD Internal Consumer Insight Report, 2024) showing that “40% of users make cross-purchases of e-commerce products on food delivery platforms.” The company is enhancing synergy between its Instant Delivery service and its core businesses to increase transaction frequency. It is also leveraging its self-operated supply chain strengths to reinforce a “quality + speed” value proposition, focusing on high–average-order-value categories such as supermarkets and 3C electronics. By reusing existing trunk logistics lines, JD.com aims to reduce costs, attract merchants through zero-commission policies, and offset subsidy pressures with high-value orders.
Regulators: Will continue to monitor market dynamics, refining policies focused on prohibiting extreme subsidies and emphasizing labor compliance. The goal is to guide the industry toward sustainable development that balances quality, efficiency, and social responsibility.
Life Science & Healthcare
Innovative drug development welcomes a new turning point
In 2025, China’s innovative drug industry is reaching a critical turning point. A series of innovation-supportive policies have been issued and implemented. In July, the National Healthcare Security Administration (NHSA) released new regulations for adjustments to the National Reimbursement Drug List (NRDL) and optimized rules for national volume-based procurement (VBP). These changes established a new “basic medical insurance + commercial insurance” dual-channel payment system and introduced several mechanisms favorable to innovative drugs—particularly in pricing, access protection periods, and clinical usage. A clear signal has been sent from the highest national level, offering long-term, full value chain support for pharmaceutical innovation. This marks a shift in development priorities from cost control to clinical value orientation and sustainable industry growth, thereby laying the foundation for high-quality development.
Building a tiered access system for healthcare payment
The core of this year’s NRDL adjustment rules lies in tiered access based on innovation value and price, with the first-time introduction of a “Commercial Insurance Innovative Drug List.” This creates a structure of “basic medical insurance covering essential needs, and commercial insurance covering high-end needs.” The basic NRDL will cover drugs of high clinical value while dynamically removing low-value, highly substitutable ones. The Commercial Insurance Innovative Drug List will target drugs that are highly innovative and high-priced, such as CAR-T and gene therapies, which will be paid for by commercial insurance, with the NHSA leading the system design and establishing the settlement platform. These measures will lower the market entry barriers for innovative drugs at the hospital level, provide differentiated pathways for innovative drugs of varying degrees of innovation, and reduce uncertainty in market access.
New VBP rules reduce price cliff risks for innovative drugs
The optimization of national volume-based procurement (VBP) rules signals a shift toward price rationalization and quality orientation. For the first time, it is explicitly stipulated that drugs within the NRDL negotiation period are excluded from VBP. This ensures price stability for three to five years after NRDL inclusion and prevents the disruptive effect of “double price cuts.”
The new bidding mechanism introduces a low-price declaration system, requiring the lowest-price bidder to submit cost analyses and commit not to bid below cost—aimed at curbing vicious competition. In parallel, the brand-based volume declaration mechanism allows hospitals to declare procurement volumes by brand, reflecting actual clinical demand and benefiting originator drugs supported by strong real-world evidence.
Quality thresholds have also been raised: successful bidders must now have at least two years of production experience and no GMP violations within the period, ensuring both supply stability and product quality.
Reshaping the innovative drug value chain to enhance sustainability
These two policies essentially lay the institutional foundation for “reshaping the value chain” of the innovative drug industry. By extending the commercial protection period for high-value products, introducing tiered access, and implementing brand-based volume declarations, the policies enable clinical value to receive more stable returns in the market, achieving a balance between “price rationalization” and “value realization.”
Figure: Full life-cycle value chain of China’s current innovative drugs
Note: 1. Depends on the level of innovation, degree of differentiation, and other lifecycle management strategies for innovative assets, such as patent protection systems
From a value chain perspective, the reforms will enable highly innovative drugs to enjoy a more sufficient market cultivation period during the introduction stages, accelerating the accumulation of clinical evidence and real-world data. In the mature stage, policy protection periods and stable payment mechanisms will slow downward price pressures, allowing companies to realize profit over a longer period. In the recession stage, dynamic NRDL delisting and competitive mechanisms will ensure market resources shift towards to R&D of new high-value assets, forming a positive cycle of continuous innovation.
Over the next three to five years, competition in China’s innovative drug industry will shift from “exchanging price for volume” to “determining volume based on value.” Companies with original innovation, clinical validation, and commercial transformation capabilities will stand out under the new policy and market environment. Greater policy certainty will also enhance predictability for investors, accelerating the deployment of capital from government, industry, and insurance. The innovative drug industry will progressively move toward a sustainable growth path that emphasizes both quality and value.
